INDEX

Explanations

concepts related to metadata or analysis

oai_token-act-pair · gpt-3.5-turbo

the term "meta" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Interstitial

-0.70

shire

-0.67

Sev

-0.65

iens

-0.64

ain

-0.63

 Tripoli

-0.63

 Opportun

-0.63

 Liberties

-0.62

HCR

-0.62

Pip

-0.61

POSITIVE LOGITS

 meta

1.03

 guiActiveUn

0.97

ethical

0.93

meta

0.89

acter

0.87

alyses

0.86

plot

0.84

allery

0.81

 referen

0.81

allo

0.79

Activations Density 0.007%

concepts related to metadata or analysis

the term "meta" in various contexts

No Comments

No Known Activations